MPC-Based Fatigue Load Suppression of Waked Wind Farm With 2Dof WT Control Strategy

Abstract

The stochastic characteristics of wind speed leads to the increase of fatigue load, and the complex wake effect makes the wind speed more complicated. To address this problem, this paper proposes an optimal wind turbine coordinated control strategy based on predicted wind speed to optimize fatigue load. Based on historical wind speed, a least square support vector machine method is proposed for future incoming wind speed prediction. A dynamic wake model is presented to quickly predict the time-varying wake wind speed changes under the change of WT mechanical state. Compared with traditional direct power control, a two-degree-of-freedom WT control strategy is proposed to jointly adjust the generator torque and pitch angle to increase the control flexibility of fatigue load suppression problem. A model predictive based control strategy considering the wind speed fluctuation is proposed to predict the future dynamic behavior of fatigue load under wind speed changes in a long prediction horizon, aiming at minimizing fatigue load and tracking power reference from transmission system operator. Wind farm simulation verifies that the proposed control scheme possesses high wind speed prediction accuracy and can significantly reduce the fatigue loads. The proposed dynamic wake model is inexpensive computationally cost for real-time wake optimization.